This guide shows you how to install and use the SLIM-CLI MCP server with Claude Code.
The SLIM CLI MCP server requires FastMCP to run. The server will attempt to install it automatically, but may require manual installation depending on your system configuration:
- Python 3.9+
- FastMCP (installed automatically or manually via
pipx install fastmcp)
- Install the MCP server using Claude Code CLI:
# From your SLIM CLI installation directory
claude mcp add slim-cli $(pwd)/src/jpl/slim/mcp/scripts/start_mcp_server.sh -s user- Verify the installation:
claude mcp list
# Should show: slim-cli - ✓ Connected- Start using SLIM in Claude Code - the server will automatically install FastMCP on first run if needed.
If you prefer manual configuration, add this to your Claude Code MCP configuration:
{
"mcpServers": {
"slim-cli": {
"command": "/path/to/slim-cli/src/jpl/slim/mcp/scripts/start_mcp_server.sh"
}
}
}Once installed, you can test the MCP server with these natural language commands in Claude Code:
"Show me all available SLIM best practices"
"List documentation-related best practices"
"What security practices are available?"
"Apply the README best practice to https://github.com/user/repo"
"Apply contributing guidelines and changelog to my repository"
"Apply the README practice with AI customization using Claude"
"What are the best AI models for documentation generation?"
"Recommend premium AI models for code generation"
"Show me local AI models I can use"
"Check if the OpenAI GPT-4 model is properly configured"
"Validate the Anthropic Claude model"
"Deploy the applied best practices to my repository"
"Push the README and contributing files to the main branch"
The MCP server provides these tools:
- slim_apply: Apply best practices to repositories
- slim_deploy: Deploy changes to git repositories
- slim_list: List available best practices
- slim_models_recommend: Get AI model recommendations
- slim_models_validate: Validate AI model configuration
- slim://registry: Complete SLIM registry data
- slim://models: AI model information and recommendations
- apply_best_practice: Template for applying practices
- review_practices: Template for reviewing practices
"I'm starting a new open source Python project. What SLIM best practices should I apply and how?"
"Help me apply documentation best practices to my repository at https://github.com/user/my-project using AI"
"Show me security-related best practices and apply them to my local repository"
"Apply governance best practices for a small team to my project"
To verify the MCP server is working:
- Check server status in Claude Code MCP settings
- Try a simple command: "List all SLIM best practices"
- Look for the tools in Claude Code's available tools list
- Name: SLIM-CLI MCP Server
- Version: 1.0.0
- Protocol: Model Context Protocol (MCP)
- Framework: FastMCP
- Python Version: 3.12+
-
"FastMCP not available"
- Make sure the virtual environment is activated
- Install FastMCP:
pip install fastmcp
-
"SLIM-CLI modules not found"
- Ensure you're running from the SLIM-CLI directory
- Check the PYTHONPATH in the configuration
-
"Server not responding"
- Restart Claude Code
- Check the MCP server logs
- Verify the configuration file syntax
- Check the server logs for error messages
- Verify all dependencies are installed
- Ensure the server file path is correct in the configuration
Once installed, you can use natural language to interact with SLIM best practices through Claude Code. The MCP server will handle the technical details while providing a conversational interface for applying, deploying, and managing SLIM best practices.
Example conversation:
You: "Apply the README and contributing guidelines to my repository using AI"
Claude: "I'll help you apply those SLIM best practices. Let me use the SLIM MCP server to apply the README and contributing practices with AI customization..."
The server makes SLIM best practices more accessible and user-friendly through natural language interaction!